使用python的tensorboard出现问题

Question

from torch.utils.tensorboard import SummaryWriter

writer=SummaryWriter("logs")
writer.add_image("test")
for i in range(100):
    writer.add_scalar("y=x",i,i)

writer.close()

上述代码报错
PS D:\zzzz\自学> & D:/ananconda/envs/pytorch1/python.exe d:/zzzz/自学/pytorch-tutorial-master/src/P8_Tensorboard.py
Traceback (most recent call last):
File "d:/zzzz/自学/pytorch-tutorial-master/src/P8_Tensorboard.py", line 1, in
from torch.utils.tensorboard import SummaryWriter
File "D:\ananconda\envs\pytorch1\lib\site-packages\torch\utils\tensorboard_init.py", line 13, in
from .writer import FileWriter, SummaryWriter # noqa: F401
File "D:\ananconda\envs\pytorch1\lib\site-packages\torch\utils\tensorboard\writer.py", line 13, in
from tensorboard.summary.writer.event_file_writer import EventFileWriter
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\summary_init.py", line 22, in
from tensorboard.summary import v1 # noqa: F401
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\summary\v1.py", line 23, in
from tensorboard.plugins.histogram import summary as histogram_summary
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\plugins\histogram\summary.py", line 35, in
from tensorboard.plugins.histogram import summary_v2
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\plugins\histogram\summary_v2.py", line 35, in
from tensorboard.util import tensor_util
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\util\tensor_util.py", line 20, in
from tensorboard.compat.tensorflow_stub import dtypes, compat, tensor_shape
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\compat\tensorflow_stub_init.py", line 25, in
from . import app # noqa
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\compat\tensorflow_stub\app.py", line 21, in
from . import flags
File "D:\ananconda\envs\pytorch1\lib\site-packages\tensorboard\compat\tensorflow_stub\flags.py", line 25, in
from absl.flags import * # pylint: disable=wildcard-import
File "D:\ananconda\envs\pytorch1\lib\site-packages\absl\flags_init_.py", line 35, in
from absl.flags import _argument_parser
File "D:\ananconda\envs\pytorch1\lib\site-packages\absl\flags_argument_parser.py", line 82, in
class ArgumentParser(Generic[_T], metaclass=_ArgumentParserCache):
TypeError: metaclass conflict: the metaclass of a derived class must be a (non-strict) subclass of the metaclasses of all its bases

Answer 1

以下回复参考：皆我百晓生、券券喵儿等免费微信小程序作答:

根据你提供的错误信息，这个问题看起来是因为你正在使用的TensorBoard库和你的PyTorch环境中的某些其他库之间存在冲突，特别是在与absl库的元类（metaclass）部分出现了冲突。错误消息表示元类冲突：派生类的元类必须是非严格的所有基类的元类的子类。这通常意味着你的环境中的库版本之间存在兼容性问题。

为了解决这个问题，你可以尝试以下几种方法：

1. 更新你的TensorBoard和absl库：运行以下命令来更新这些库，确保它们是兼容的版本。

pip install --upgrade tensorboard absl-py

这可能会解决库之间的兼容性问题。

1. 创建一个新的虚拟环境并安装所需的库：有时，库之间的冲突可能是由于在不同的环境中安装的库版本不同造成的。创建一个新的Python虚拟环境，并在其中安装TensorBoard和PyTorch。这样可以帮助确保所有的库都在一个独立的、一致的环境中安装。

例如，使用conda创建新环境并安装所需的库：

conda create -n new_env python=3.8
conda activate new_env
pip install torch tensorboard absl-py

然后在新环境中运行你的代码。

1. 检查你的代码中的导入语句：确保你没有错误地导入任何库或模块。根据你提供的代码片段，导入语句看起来是正确的，但确保你安装了正确的库版本并正确引用了它们。

如果上述方法都不能解决问题，可能需要考虑查看TensorBoard的官方文档或相关社区，看看是否有其他用户报告了类似的问题和解决方案。